{
  "status": "research",
  "advantages": [],
  "limitations": [],
  "supersedes": [],
  "supersededBy": [],
  "firstSeen": null,
  "lastSeen": null,
  "updated": "2026-11-20",
  "slug": "xfetch",
  "term": "XFetch",
  "aliases": [
    "Exponential early expiration",
    "Probabilistic early expiration (XFetch)"
  ],
  "definition": "XFetch is a cache-refresh algorithm that uses an exponentially distributed time gap to trigger recomputation before a cached value expires.",
  "note": "Mechanism: A request refreshes when a random look-ahead exceeds the remaining lifetime. The gap scale uses the previous recomputation duration and a positive tuning factor. Larger scales favor earlier refresh. [xfetch-paper]\n\nScope: The paper's bounds concern an arrival model and expected behavior, not a one-refresher guarantee for every request burst. [xfetch-paper]",
  "related": [
    "cache-stampede",
    "request-coalescing",
    "cache-admission"
  ],
  "citations": [
    {
      "id": "xfetch-paper",
      "title": "Vattani, Chierichetti and Lowenstein. Optimal Probabilistic Cache Stampede Prevention. PVLDB 8(8), 886-897 (2015), sections 2-5.",
      "url": "https://cseweb.ucsd.edu/~avattani/papers/cache_stampede.pdf",
      "accessed": "2026-11-20"
    }
  ],
  "provenance": [
    "Section 5 and Figure 3 supply the implementation rule; sections 2-4 define the renewal-arrival model and asymptotic bounds. No claim of workload replication."
  ]
}
